US5982669A

US5982669A - EPROM and flash memory cells with source-side injection

Info

Publication number: US5982669A
Application number: US09/185,892
Authority: US
Inventors: Alexander Kalnitsky; Christoph Pichler; Albert Bergemont
Original assignee: National Semiconductor Corp
Current assignee: National Semiconductor Corp
Priority date: 1998-11-04
Filing date: 1998-11-04
Publication date: 1999-11-09
Anticipated expiration: 2018-11-04

Abstract

An electrically-programmable read-only-memory (EPROM) and a flash memory cell provide both source-side and drain-side injection, along with a reduced cell size, by forming the memory cell in a trench. The drain is formed in the top surface of the substrate, the source is formed in the bottom surface of the trench, and the stacked gate is formed over the sidewall of the trench.

Description

RELATED APPLICATION

The present application is related to application Ser. No. 09/185,893 for Method of Forming EPROM and Flash Memory Cells with Source-Side Injection by Alexander Kalnitsky et al. which is filed on an even date herewith.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to EPROM and flash memory cells and, more particularly, EPROM and flash memory cells with source-side injection.

2. Description of the Related Art

An electrically-programmable read-only-memory (EPROM) cell and a flash memory cell are non-volatile memories that retain data stored in the cell after power to the cell has been removed. EPROM and flash memory cells principally differ from each other in that EPROM cells are erased with ultraviolet (UV) light, while flash cells are electrically erased.

FIG. 1 shows a cross-sectional view that illustrates a prior art EPROM or flash memory cell 100. As shown in FIG. 1, cell 100 includes spaced-apart n+ source and

drain regions

112 and 114 which are formed in a p-type substrate 110, and a channel region 116 which is defined in substrate 110 between source and

drain regions

112 and 114.

In addition, cell 100 also includes a layer of gate oxide 120 which is formed over channel region 116, and a floating gate 122 which is formed over gate oxide layer 120. Further, cell 100 additionally includes a layer of interpoly dielectric 124 which is formed over floating gate 122, and a control gate 126 which is formed over dielectric layer 124.

Cell 100 is programmed by applying a programming voltage to control gate 126, a drain voltage to drain region 114, and ground to source region 112. The programming voltage applied to control gate 126 induces a positive potential on floating gate 122 which, in turn, attracts electrons to the surface of channel region 116 to form a channel 130.

In addition, the source-to-drain voltage sets up an electric field which causes electrons to flow from source region 112 to drain region 114 via channel 130. As the electrons flow to drain region 114, the electric field, which has a maximum near drain region 114, accelerates these electrons into having ionizing collisions that form channel hot electrons near drain region 114.

A small percentage of the channel hot electrons are then injected onto floating gate 122 via gate oxide layer 120. Cell 100 is programmed when the number of electrons injected onto floating gate 122 is sufficient to prevent channel 130 from being formed when a read voltage is subsequently applied to control gate 124.

Since only a small percentage of the channel hot electrons are injected onto floating gate 122, channel hot electron programming provides a relatively low injection efficiency. One technique for increasing the injection efficiency is to create a source-to-drain electric field which, in addition to having a peak near the drain region, also has a peak near the source region. EPROM and flash memory cells which utilize an electric field which has a peak near the source region are typically referred to as having source side injection.

One way of forming an electric field which has a peak near both the source and drain regions, as described in U.S. Pat. No. 4,652,897 to Okuyama et al., is to use a low-density source region which is adjacent to the source region.

FIG. 2 shows a cross-sectional view that illustrates a prior-art, source-side injection EPROM or flash memory cell 200. FIG. 2 is similar to FIG. 1 and, as a result, utilizes the same reference numerals to designate the structures which are common to both cells.

As shown in FIG. 2, source-side cell 200 differs from cell 100 in that cell 200 also includes a n-type low-density source (LDS) region 210 which is formed between source region 112 and channel region 130. LDS region 210 has a dopant concentration which is less than the dopant concentration of source region 112.

In operation, cell 200 is programmed the same as cell 100 except that the presence of LDS region 210 causes the source-to-drain electric field to have peaks near both the LDS and drain regions 210 and 114. As a result, channel hot electrons are formed and injected onto floating gate 122 near both of these regions 210 and 114, thereby increasing the injection efficiency.

FIG. 3 shows a graph that illustrates the intensity of the electric field along the surface of the channel region between the drain and source regions of cell 200. As shown in FIG. 3, line L0 illustrates that cell 200 has a peak in the intensity of the electric field near both the drain and source regions.

One disadvantage of cell 200, however, is that as a result of LDS region 210, cell 200 consumes significantly more silicon real estate than does cell 100. Thus, there is a need for an EPROM or flash memory cell that provides an increased injection efficiency with reduced cell size.

SUMMARY OF THE INVENTION

The present invention provides an electrically-programmable read-only-memory (EPROM) or a flash memory cell with both source-side and drain-side injection, along with a reduced cell size, by forming the memory cell along the side of a trench.

The memory cell of the present invention includes a semiconductor material of a first conductivity type. The material has a first surface, a second surface vertically spaced-apart from and below the first surface, and a sidewall surface which is connected to the first surface at point A, and to the second surface at point B. A line drawn through points A and B, and a line normal to the first surface through point A define an angle which ranges from greater than zero to less than 90 degrees.

The cell of the present invention also includes a source region of a second conductivity type which is formed in the material to contact the second surface, and a drain region of the second conductivity type which is formed in the material to contact the first surface.

The cell of the present invention further includes a channel region which is defined in the material between the source and drain regions, and a layer of gate oxide which is formed on the material over the channel region. In addition, a floating gate is formed over the layer of gate oxide, a layer of interpoly dielectric is formed over the floating gate, and a control gate is formed over the layer of interpoly dielectric.

A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description and accompanying drawings which set forth an illustrative embodiment in which the principals of the invention are utilized.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a cross-sectional view illustrating a prior art EPROM or flash memory cell 100.

FIG. 2 is a cross-sectional view illustrating a prior-art, drain and source-side injection EPROM or flash memory cell 200.

FIG. 3 is a graph illustrating the intensity of the electric field along the surface of the channel region between the drain and source regions of cell 200.

FIG. 4 is a cross-sectional view illustrating an EPROM or flash EPROM cell 400 in accordance with the present invention.

FIG. 5 is a graph illustrating the intensity of an electric field along the surface of the channel region between the drain and source regions of

cells

100 and 400 in accordance with the present invention.

FIG. 6 is a schematic diagram illustrating a segment 600 of a plurality of segments that form a memory array in accordance with the present invention.

FIGS. 7A-11A are plan views illustrating a process for forming segment 600 in accordance with the present invention.

FIGS. 7B-11B are cross-sectional views taken along lines 7B--7B, 8B--8B, 9B--9B, 10B--10B, and 11B--11B of FIGS. 7A, 8A, 9A, 10A, and 11A, respectively.

FIGS. 7C-11C are cross-sectional views taken along lines 7C--7C, 8C--8C, 9C--9C, 10C--10C, and 11C--11C of FIGS. 7A, 8A, 9A, 10A, and 11A, respectively.

DETAILED DESCRIPTION

FIG. 4 shows a cross-sectional view that illustrates an EPROM or flash EPROM cell 400 in accordance with the present invention. As described in greater detail below, cell 400 provides both source-side and drain-side injection, along with a reduced cell size, by forming cell 400 along the sidewall of a trench.

As shown in FIG. 4, cell 400 includes a p-type semiconductor material 410, such as a substrate or a well. Material 410, in turn, has a first surface 412, a second surface 414 which is vertically and horizontally spaced-apart from first surface 412, and a sidewall surface 416 which is connected to first surface 412 at point A, and to second surface 414 at point B. A line drawn through points A and B, and a line normal to first surface 412 through point A define an acute angle φ which ranges from greater than zero to less than 90 degrees.

As further shown in FIG. 4, cell 400 includes a n+source region 420 which is formed in material 410 to contact second surface 414, and a n+ drain region 422 which is formed in material 410 to contact first surface 412. Cell 400 also includes a channel region 424 which is defined in material 410 between source and drain

regions

420 and 422, and a layer of gate oxide 426 which is formed on material 410 over channel region 424.

In addition, cell 400 further includes a floating gate 430 which is formed over gate oxide layer 426, a layer of interpoly dielectric 432 which is formed over floating gate 430, and a control gate 434 which is formed over dielectric layer 432.

In operation, cell 400 is programmed the same as cell 100 except that the geometry of the cell causes the source-to-drain electric field to have peaks near both the source and drain

regions

420 and 422. As a result, channel hot electrons are formed and injected onto floating gate 430 near both of these

regions

420 and 422.

FIG. 5 shows a graph that illustrates the drain-to-source electric field intensity along the surface of the channel region between the drain and source regions of

cells

100 and 400 during programming. As shown in FIG. 5, line L1 illustrates the electric field intensity in cell 100 of FIG. 1, while line L2 illustrates the electric field intensity in cell 400 (where

cells

100 and 400 have approximately the same gate length and gate oxide thickness).

As shown in FIG. 5,

cells

100 and 400 both have peaks near the drain region. Near the source region, however, cell 100 has a relatively flat field intensity whereas cell 400 has a second peak which is approximately 1.5 to 2.0 times greater than the field intensity of cell 100.

As a result, cell 400 has a greater injection efficiency than does cell 100. One of the advantages of having a greater injection efficiency is that cell 400 can be programmed in less time than is required to program cell 100.

FIG. 6 shows a schematic diagram that illustrates a segment 600 of a plurality of segments that form a memory array in accordance with the present invention. (Rather than representing a segment, FIG. 6 may also be viewed as representing the entire memory array).

As shown in FIG. 6, segment 600 includes a plurality of cells 400 (of FIG. 4) which are arranged in rows and columns. Segment 600 also includes a plurality of contacted bit lines CBL1-CBLr which are connected to the drains of cells 400 such that each contacted bit line CBL is connected to all of the drains in two adjacent columns of cells. Each contacted bit line CBL, in turn, is connected to a metal bit line contact MBL. Segment 600 further includes a plurality of noncontacted bit lines NCB1-NCBs which are formed so that one non-contacted bit line NCB is positioned between each pair of adjacent contacted bit lines CBL, and connected to the sources of cells 400 such that each non-contacted bit line NCB is connected to all of the sources in two adjacent columns of cells.

Segment 600 additionally includes a plurality of access transistors 610 which are formed in four rows so that an access transistor 610 is formed at each end of each column of memory cells 112. In addition, each row of access transistors 610 is formed so that an access transistor 610 is connected to every other pair of adjacent contacted and noncontacted bit lines CBL1-CBLr and NCB1-NCBs

As further shown in FIG. 6, segment 600 includes a plurality of word lines WL1-WLn which are connected to cells 400 so that a word line WL is connected to each cell 400 in a row of cells 400. Similarly, segment 600 also includes four access lines AL1-AL4 which are connected to access transistors 610 so that an access line AL is connected to each access transistor 610 in a row of access transistors 610.

To program a selected cell in segment 600, an access voltage is applied to the access lines AL that are connected to the access transistors 610 that are connected to the source region of the selected cell, while ground is applied to the remaining access lines AL. For example, if cell A is to be programmed, then an access voltage is applied to access line AL2 and access line AL3 while ground is applied to access lines AL1 and AL4.

Applying an access voltage to access lines AL2 and AL3 turns on the access transistors in the two columns adjacent to the column that includes cell A which, in turn, electrically connects cell A to contacted bit line CBL3.

In addition, a drain voltage is applied to the contacted bit line CBL that is connected to the cell to be programmed, while ground is applied to the remaining contacted bit lines CBL. Thus, to program cell A, the drain voltage is applied to contacted bit line CBL2 which places the drain voltage on the drain of cell A. Since cell A is electrically connected to contacted bit line CBL3 via the access transistors, ground is connected to the source of cell A via the access transistors and non-contacted bit line NCB2.

Further, a programming voltage is applied to the word line WL that corresponds to the cell to be programmed, and ground to the remaining word lines WL. Thus, to program cell A, the programming voltage is applied to word line WL1 while ground is applied to the remaining word lines WL2-WLn.

These bias conditions result in a current flow as shown by the arrow in FIG. 6 which results in electrons being injected onto the floating gate of cell A near both the source and drain of cell A, thus programming cell A.

FIGS. 7A-11A show a series of plan views that illustrate a process for forming segment 600 in accordance with the present invention. FIGS. 7B-11B show a series of cross-sectional views taken along lines 7B--7B, 8B--8B, 9B--9B, 10B--10B, and 11B--11B of FIGS. 7A, 8A, 9A, 10A, and 11A, respectively. FIGS. 7C-11C show a series of cross-sectional views taken along lines 7C--7C, 8C--8C, 9C--9C, 10C--10C, and 11C--11C of FIGS. 7A, 8A, 9A, 10A, and 11A, respectively.

As shown in FIGS. 7A-7C, the process of the present invention begins with a conventionally formed semiconductor material 710, such as a substrate or well, which has a top surface 712. Following this, a trench mask 714 is formed and patterned on substrate 710.

Next, the exposed regions of substrate 710 are isotropically etched to form a series of trenches 716. (Alternately, substrate 710 may be anisotropically etched, and then isotropically etched to vary the shape of the trench). Substrate 710 is implanted to set the threshold voltages of the transistors after the trench formation.

As shown in FIGS. 7A-7C, each trench 716 has a bottom surface 720 and a pair of sidewall surfaces 722 which connect bottom surface 720 to the pair of adjoining top surfaces 712. Top surface 712 of substrate 710 represents a first surface (412) while bottom surface 720 of trench 716 represents a second surface (414).

Further, each sidewall surface 722 (416) is connected to top surface 712 at a point A, and to bottom surface 720 at a point B. A line drawn through points A and B, and a line normal to top surface 712 through point A define an angle φ which ranges from greater than zero to less than 90 degrees. After the series of trenches 716 have been formed, trench mask 714 is removed.

As shown in FIGS. 8A-8C, once mask 714 has been removed, a layer of gate oxide 724 is formed over top surface 712, bottom surface 720, and sidewall surfaces 722 of substrate 710. Next, a layer of first polysilicon (poly-1) 726 is deposited on gate oxide layer 724, and then doped.

Once poly-1 layer 726 has been doped, a layer of interpoly dielectric 730, such as ONO (oxide-nitride-oxide), is formed on poly-1 layer 726. Following this, a second layer of polysilicon (poly-2) 732 is deposited on dielectric layer 730, and then doped. Once poly-2 layer 732 has been doped, a stacked gate mask 734 is formed and patterned on poly-2 layer 732.

As shown in FIGS. 9A-9C, the exposed regions of poly-2 layer 732, and the underlying layers of interpoly dielectric and poly-1 730 and 726, are anisotropically etched to form a plurality of stacked gates structures 736 which each have a floating gate 738 and a control gate 740, and an exposed gate oxide region. (This etching step may cause some trenching of the substrate surface due to the severe topology).

As further shown in FIGS. 9A-9C, access transistors 610 will be formed from the stacked gate structures 736 in the first two rows, while memory transistors 400 will be formed from the stacked gate structures 736 in the third and fourth rows.

Following this, mask 734 is removed. A light dose of p-type material, such as boron, is then implanted using a large angle tilt or plasma immersion techniques to form channel stop regions 741 between each stacked gate structure 736 in each column of stack gate structures 736. After this, the wafer is annealed in an oxidizing ambient to repair etch damage and drive in the channel stop implant.

Next, an implant mask 742 is then formed and patterned over the stacked gate structures 736 and portions of the exposed gate oxide region to define a plurality of implant strips 744 on top and

bottom surfaces

712 and 720. Next, a n-type material, such as arsenic or phosphorous, is implanted into the implant strips 744 to form n+ source and

drain lines

746 and 750. After this, mask 742 is removed.

As shown in FIGS. 10A-10C, once the source and

drain lines

746 and 750 have been formed, a layer of oxide dielectric 754 is deposited over stacked gate structures 736, source and drain

regions

746 and 750, and isolation regions 752, and then planarized. Dielectric layer 754 serves both as the poly-to-metal isolation and the field oxide isolation.

Next, a contact opening mask 756 is formed and patterned on dielectric layer 754 to define a contact region 760 on the surface of each control gate 740. Next, the exposed regions of dielectric layer 754 are anisotropically etched until dielectric layer 754 has been removed from contact region 760 on the surface of each control gate 740. Mask 756 is then removed. Once mask 756 has been stripped, a layer of contact material (not shown) is deposited on dielectric layer 754 and the contact regions 760. As shown in FIGS. 11A-11C, the contact material is then etched until the contact material has been removed from the top surface of dielectric layer 754, thereby forming contacts 762. Once contacts 762 have been formed, a layer of metal (not shown) is deposited over dielectric layer 754 and contacts 762. Following this, an access/word line mask (not shown) is formed and patterned over the layer of metal.

Next, the exposed regions of the layer of metal are anisotropically etched until the layer of metal has been removed from dielectric layer 754, thereby forming access lines 764 (AL1-AL4) and word lines 766 (WL1-WLn). The access/word line mask is then removed, and the process continues with conventional steps.

As described in the above process flow, access transistors 610 are formed to be stacked-gate transistors. Access transistors 610 may alternately be formed to be single-poly transistors during the same process steps that are used to form the n-channel and p-channel peripheral transistors. When access transistors 610 are formed to be stacked gate transistors, the access transistors are formed to be larger than memory transistors 400 to increase the charge flow through the access transistors 610.

It should be understood that various alternatives to the embodiment of the invention described herein may be employed in practicing the invention. Thus, it is intended that the following claims define the scope of the invention and that methods and structures within the scope of these claims and their equivalents be covered thereby.

Claims

What is claimed is:

1. A memory device comprising:

a plurality of memory cells arranged in rows and columns, each cell having a drain region, a source region vertically and horizontally spaced-apart from the drain region, a channel region defined between the drain and source regions, a floating gate formed over the channel region, and a control gate formed over the floating gate, the floating gate and the control gate each having a first end and a second end which is vertically and horizontally spaced-apart from the first end;

a plurality of contacted bit lines connected to the drains of the cells such that each contacted bit line is connected to all of the drains in two adjacent columns of cells;

a plurality of non-contacted bit lines formed so that one non-contacted bit line is positioned between each pair of adjacent contacted bit lines, and connected to the sources of the cells such that each non-contacted bit line is connected to all of the sources in two adjacent columns of cells;

a plurality of access transistors formed in four rows so that an access transistor is formed at each end of each column of memory cells, each row of access transistors being formed so that an access transistor is connected to every other pair of adjacent contacted and noncontacted bit lines;

a plurality of word lines connected to the cells so that a word line is connected to each cell in a row of cells;

a plurality of access lines connected to the access transistors so that an access line is connected to each access transistor in a row of access transistors.

2. A method for programming a selected memory cell in a memory device, the memory device having:

a plurality of memory cells arranged in rows and columns, each cell having a drain region, a source region vertically spaced-apart from and below the drain region, a channel region defined between the drain and source regions, a floating gate formed over the channel region, and a control gate formed over the floating gate, the floating gate and the control gate each having a first end and a second end which is vertically spaced-apart from and below the first end;

a plurality of contacted bit lines connected to the drains of the cells such that each contacted bit line is connected to all of the drains in two adjacent columns of cells, the plurality of contacted bit lines including a contacted bit line connected to the drain of the selected memory cell and a plurality of remaining contacted bit lines;

a plurality of word lines connected to the cells so that a word line is connected to each cell in a row of cells, the plurality of word lines including a selected word line connected to the control gate of the selected memory cell and a plurality of remaining word lines;

a plurality of access lines connected to the access transistors so that an access line is connected to each access transistor in a row of access transistors, the plurality of access lines including a pair of selected access lines connected to access transistors connected to the source of the selected memory cell and unconnected to the drain of the selected memory cell, and a pair of remaining access lines,

the method comprising the steps of:

applying an access voltage to the pair of selected access lines;

applying ground to the pair of remaining access lines;

applying a drain voltage to the contacted bit line that is connected to the selected memory cell;

applying ground to the remaining contacted bit lines;

applying a programming voltage to the word line connected to the selected memory cell; and

applying ground to the remaining word lines.

3. The memory device of claim 1 wherein a word line is connected to each control gate in a row of cells.

4. A memory device comprising:

a plurality of access transistors formed in a plurality of rows so that an access transistor is formed at each end of each column of memory cells, each row of access transistors being formed so that an access transistor is connected to every other pair of adjacent contacted and noncontacted bit lines;

5. The memory device of claim 4 wherein a word line is connected to each control gate in a row of cells.

6. A method for programming a selected memory cell in a memory device, the memory device having:

a plurality of access lines connected to the access transistors so that an access line is connected to each access transistor in a row of access transistors,

the method comprising the steps of:

applying an access voltage to each access line connected to an access transistor that is connected to the source of the selected cell and unconnected to the drain of the selected cell;

applying ground to each access line unconnected to an access transistor that is connected to the source of the selected cell and unconnected to the drain of the selected cell;

applying ground to the remaining contacted bit lines;

applying ground to the remaining word lines.

7. The method of claim 4 wherein a word line is connected to each control gate in a row of cells.